12 KiB
12 KiB
AITech — Uczenie maszynowe — laboratoria
11. Sieci neuronowe (Keras)
Keras to napisany w języku Python interfejs do platformy TensorFlow, służącej do uczenia maszynowego.
Aby z niego korzystać, trzeba zainstalować bibliotekę TensorFlow:
Przykład implementacji sieci neuronowej do rozpoznawania cyfr ze zbioru MNIST, według https://keras.io/examples/vision/mnist_convnet
# Konieczne importy
import numpy as np
from tensorflow import keras
from tensorflow.keras import layers2023-06-01 10:29:41.492705: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2 FMA To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags. 2023-06-01 10:29:42.477407: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudart.so.11.0'; dlerror: libcudart.so.11.0: cannot open shared object file: No such file or directory 2023-06-01 10:29:42.477524: I tensorflow/compiler/xla/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine. 2023-06-01 10:29:45.603958: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory 2023-06-01 10:29:45.604816: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory 2023-06-01 10:29:45.604834: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
# Przygotowanie danych
num_classes = 10
input_shape = (28, 28, 1)
# podział danych na zbiory uczący i testowy
(x_train, y_train), (x_test, y_test) = keras.datasets.mnist.load_data()
# skalowanie obrazów do przedziału [0, 1]
x_train = x_train.astype("float32") / 255
x_test = x_test.astype("float32") / 255
# upewnienie się, że obrazy mają wymiary (28, 28, 1)
x_train = np.expand_dims(x_train, -1)
x_test = np.expand_dims(x_test, -1)
print("x_train shape:", x_train.shape)
print(x_train.shape[0], "train samples")
print(x_test.shape[0], "test samples")
# konwersja danych kategorycznych na binarne
y_train = keras.utils.to_categorical(y_train, num_classes)
y_test = keras.utils.to_categorical(y_test, num_classes)x_train shape: (60000, 28, 28, 1) 60000 train samples 10000 test samples
# Stworzenie modelu
model = keras.Sequential(
[
keras.Input(shape=input_shape),
layers.Conv2D(32, kernel_size=(3, 3), activation="relu"),
layers.MaxPooling2D(pool_size=(2, 2)),
layers.Conv2D(64, kernel_size=(3, 3), activation="relu"),
layers.MaxPooling2D(pool_size=(2, 2)),
layers.Flatten(),
layers.Dropout(0.5),
layers.Dense(num_classes, activation="softmax"),
]
)
model.summary()Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d (Conv2D) (None, 26, 26, 32) 320
max_pooling2d (MaxPooling2D (None, 13, 13, 32) 0
)
conv2d_1 (Conv2D) (None, 11, 11, 64) 18496
2023-06-01 10:29:49.494604: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory 2023-06-01 10:29:49.495467: W tensorflow/compiler/xla/stream_executor/cuda/cuda_driver.cc:265] failed call to cuInit: UNKNOWN ERROR (303) 2023-06-01 10:29:49.496113: I tensorflow/compiler/xla/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (ELLIOT): /proc/driver/nvidia/version does not exist 2023-06-01 10:29:49.497742: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2 FMA To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
max_pooling2d_1 (MaxPooling (None, 5, 5, 64) 0
2D)
flatten (Flatten) (None, 1600) 0
dropout (Dropout) (None, 1600) 0
dense (Dense) (None, 10) 16010
=================================================================
Total params: 34,826
Trainable params: 34,826
Non-trainable params: 0
_________________________________________________________________
# Uczenie modelu
model.compile(loss="categorical_crossentropy", optimizer="adam", metrics=["accuracy"])
model.fit(x_train, y_train, batch_size=128, epochs=15, validation_split=0.1)2023-06-01 10:30:24.247916: W tensorflow/tsl/framework/cpu_allocator_impl.cc:82] Allocation of 169344000 exceeds 10% of free system memory.
Epoch 1/15 422/422 [==============================] - 36s 82ms/step - loss: 0.3806 - accuracy: 0.8831 - val_loss: 0.0894 - val_accuracy: 0.9738 Epoch 2/15 422/422 [==============================] - 34s 80ms/step - loss: 0.1174 - accuracy: 0.9644 - val_loss: 0.0611 - val_accuracy: 0.9827 Epoch 3/15 422/422 [==============================] - 63s 149ms/step - loss: 0.0858 - accuracy: 0.9739 - val_loss: 0.0482 - val_accuracy: 0.9870 Epoch 4/15 422/422 [==============================] - 29s 68ms/step - loss: 0.0748 - accuracy: 0.9762 - val_loss: 0.0431 - val_accuracy: 0.9885 Epoch 5/15 422/422 [==============================] - 35s 84ms/step - loss: 0.0644 - accuracy: 0.9804 - val_loss: 0.0391 - val_accuracy: 0.9898 Epoch 6/15 422/422 [==============================] - 32s 75ms/step - loss: 0.0562 - accuracy: 0.9826 - val_loss: 0.0367 - val_accuracy: 0.9908 Epoch 7/15 422/422 [==============================] - 29s 68ms/step - loss: 0.0521 - accuracy: 0.9841 - val_loss: 0.0356 - val_accuracy: 0.9897 Epoch 8/15 422/422 [==============================] - 28s 67ms/step - loss: 0.0484 - accuracy: 0.9840 - val_loss: 0.0334 - val_accuracy: 0.9922 Epoch 9/15 422/422 [==============================] - 29s 69ms/step - loss: 0.0466 - accuracy: 0.9855 - val_loss: 0.0355 - val_accuracy: 0.9908 Epoch 10/15 422/422 [==============================] - 29s 68ms/step - loss: 0.0423 - accuracy: 0.9864 - val_loss: 0.0332 - val_accuracy: 0.9902 Epoch 11/15 422/422 [==============================] - 30s 71ms/step - loss: 0.0413 - accuracy: 0.9868 - val_loss: 0.0315 - val_accuracy: 0.9915 Epoch 12/15 422/422 [==============================] - 29s 68ms/step - loss: 0.0380 - accuracy: 0.9876 - val_loss: 0.0294 - val_accuracy: 0.9913 Epoch 13/15 422/422 [==============================] - 30s 70ms/step - loss: 0.0371 - accuracy: 0.9883 - val_loss: 0.0287 - val_accuracy: 0.9917 Epoch 14/15 422/422 [==============================] - 29s 70ms/step - loss: 0.0342 - accuracy: 0.9886 - val_loss: 0.0380 - val_accuracy: 0.9893 Epoch 15/15 422/422 [==============================] - 29s 68ms/step - loss: 0.0351 - accuracy: 0.9888 - val_loss: 0.0320 - val_accuracy: 0.9912
<keras.callbacks.History at 0x7f50553cc760>
# Ewaluacja modelu
score = model.evaluate(x_test, y_test, verbose=0)
print("Test loss:", score[0])
print("Test accuracy:", score[1])